In [31]:
import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score, mean_squared_error, mean_absolute_error
import matplotlib.pyplot as plt
# List to store models and metrics
train_r2_scores = []
validation_r2_scores = []
train_mse_scores = []
validation_mse_scores = []
train_rmse_scores = []
validation_rmse_scores = []
train_mae_scores = []
validation_mae_scores = []
# List to store baseline metrics
baseline_r2_scores = []
baseline_mse_scores = []
baseline_rmse_scores = []
baseline_mae_scores = []
port_values = []
# Iterate through each row in the DataFrame
for index, row in df.iterrows():
port_value = row['port'] # Get 'port' column value for the current row
port_values.append(port_value)
# Extract features (excluding first column -- years) target -- all data points
X = np.array(row.index[1:]).reshape(-1, 1) # Features using row indices
y = np.array(row.values[1:]).reshape(-1, 1) # Target using row values
# Identify and remove outliers using IQR method
Q1 = np.percentile(y, 25)
Q3 = np.percentile(y, 75)
IQR = Q3 - Q1
lower_bound = Q1 - 1.5 * IQR
upper_bound = Q3 + 1.5 * IQR
filtered_indices = np.where((y >= lower_bound) & (y <= upper_bound))[0]
X = X[filtered_indices]
y = y[filtered_indices]
# Create polynomial features
degree = 3
poly = PolynomialFeatures(degree=degree)
X_poly = poly.fit_transform(X)
# Apply bootstrapping to generate more data
num_bootstrap_samples = 5 # Number of bootstrap samples
bootstrap_X = []
bootstrap_y = []
for _ in range(num_bootstrap_samples):
# Resample with replacement
bootstrap_indices = np.random.choice(len(X_poly), size=len(X_poly), replace=True)
bootstrap_X.append(X_poly[bootstrap_indices])
bootstrap_y.append(y[bootstrap_indices])
bootstrap_X = np.vstack(bootstrap_X)
bootstrap_y = np.vstack(bootstrap_y)
# Split data into train and validation sets
X_train, X_val, y_train, y_val = train_test_split(bootstrap_X, bootstrap_y, test_size=0.2, random_state=1)
# Train the polynomial regression model
model = LinearRegression()
model.fit(X_train, y_train)
# Make predictions on training and validation data
y_train_pred = model.predict(X_train)
y_val_pred = model.predict(X_val)
mean_prediction = np.mean(y_train) # Use mean of training data as baseline prediction
# Calculate R² score for training and validation data
train_r2 = r2_score(y_train, y_train_pred)
validation_r2 = r2_score(y_val, y_val_pred)
# Calculate MSE, RMSE, and MAE for training and validation data
train_mse = mean_squared_error(y_train, y_train_pred)
validation_mse = mean_squared_error(y_val, y_val_pred)
train_rmse = np.sqrt(train_mse)
validation_rmse = np.sqrt(validation_mse)
train_mae = mean_absolute_error(y_train, y_train_pred)
validation_mae = mean_absolute_error(y_val, y_val_pred)
# Append metrics to the respective lists
train_r2_scores.append(train_r2)
validation_r2_scores.append(validation_r2)
train_mse_scores.append(train_mse)
validation_mse_scores.append(validation_mse)
train_rmse_scores.append(train_rmse)
validation_rmse_scores.append(validation_rmse)
train_mae_scores.append(train_mae)
validation_mae_scores.append(validation_mae)
# Append baseline metrics to the respective lists
baseline_r2_scores.append(r2_score(y_val, np.full_like(y_val, mean_prediction)))
baseline_mse_scores.append(mean_squared_error(y_val, np.full_like(y_val, mean_prediction)))
baseline_rmse_scores.append(np.sqrt(mean_squared_error(y_val, np.full_like(y_val, mean_prediction))))
baseline_mae_scores.append(mean_absolute_error(y_val, np.full_like(y_val, mean_prediction)))
# Plot the polynomial curve for the entire range (1972 to 2024)
# Plot the training data
plt.figure(figsize=(8, 6))
plt.scatter(X_train[:, 1], y_train, color='blue', label='Training Data', alpha=0.7) # Training data points
sort_indices_train = np.argsort(X_train[:, 1])
plt.plot(X_train[sort_indices_train, 1], y_train_pred[sort_indices_train], color='red', linewidth=2, label='Polynomial Fit (Train)')
plt.axhline(y=mean_prediction, color='purple', linestyle='--', label='Baseline (Mean Prediction)')
plt.xlabel('Feature (Index)')
plt.ylabel('Target')
plt.title(f'Row {index + 1} Training Data and Model - Train R²: {train_r2:.2f}')
plt.legend()
plt.show()
# Plot the validation data
plt.figure(figsize=(8, 6))
plt.scatter(X_val[:, 1], y_val, color='green', label='Validation Data', alpha=0.7) # Validation data points
sort_indices_val = np.argsort(X_val[:, 1])
plt.plot(X_val[sort_indices_val, 1], y_val_pred[sort_indices_val], color='orange', linewidth=2, label='Polynomial Fit (Validation)')
plt.axhline(y=mean_prediction, color='purple', linestyle='--', label='Baseline (Mean Prediction)')
plt.xlabel('Feature (Index)')
plt.ylabel('Target')
plt.title(f'Row {index + 1} Validation Data and Model - Validation R²: {validation_r2:.2f}')
plt.legend()
plt.show()
# Plot the full range with baseline and validation points
plt.figure(figsize=(8, 6))
X_full = np.arange(1972, 2025).reshape(-1, 1)
X_full_poly = poly.transform(X_full) # Use transform instead of fit_transform
y_full_pred = model.predict(X_full_poly)
plt.scatter(X_val[:, 1], y_val, color='green', label='Validation Data', alpha=0.7) # Validation data points
plt.plot(X_full[:, 0], y_full_pred, color='purple', linewidth=2, label='Polynomial Fit (Full Range)') # Accessing the first column of X_full
plt.axhline(y=mean_prediction, color='purple', linestyle='--', label='Baseline (Mean Prediction)')
plt.xlabel('Feature (Index)')
plt.ylabel('Target')
plt.title(f'Row {index + 1} Full Range and Baseline')
plt.legend()
plt.show()
print(f"Row {index + 1} - Port Value: {port_value}")
# Print the metrics for each row's model and baseline
print(f"Row {index + 1}'s Model - Train R²: {train_r2:.2f}, Validation R²: {validation_r2:.2f}")
print(f"Train MSE: {train_mse:.2f}, Validation MSE: {validation_mse:.2f}")
print(f"Train RMSE: {train_rmse:.2f}, Validation RMSE: {validation_rmse:.2f}")
print(f"Train MAE: {train_mae:.2f}, Validation MAE: {validation_mae:.2f}")
print(f"Baseline R²: {baseline_r2_scores[-1]:.2f}, Baseline MSE: {baseline_mse_scores[-1]:.2f}, Baseline RMSE: {baseline_rmse_scores[-1]:.2f}, Baseline MAE: {baseline_mae_scores[-1]:.2f}")
print(f"Theoretical Prediction for 2024: {y_full_pred[-1]}") # Print the theoretical prediction for 2024
print("\n")
Row 1 - Port Value: Albany (NY) Row 1's Model - Train R²: 0.71, Validation R²: 0.66 Train MSE: 990781.16, Validation MSE: 901078.68 Train RMSE: 995.38, Validation RMSE: 949.25 Train MAE: 816.34, Validation MAE: 758.93 Baseline R²: -0.13, Baseline MSE: 2961999.76, Baseline RMSE: 1721.05, Baseline MAE: 1606.66 Theoretical Prediction for 2024: [190.17560482]
Row 2 - Port Value: Alpena (MI) Row 2's Model - Train R²: 0.72, Validation R²: 0.64 Train MSE: 45361.37, Validation MSE: 50935.25 Train RMSE: 212.98, Validation RMSE: 225.69 Train MAE: 164.74, Validation MAE: 190.57 Baseline R²: -0.00, Baseline MSE: 141003.39, Baseline RMSE: 375.50, Baseline MAE: 326.19 Theoretical Prediction for 2024: [18437.83449364]
Row 3 - Port Value: Anacortes (WA) Row 3's Model - Train R²: 0.90, Validation R²: 0.80 Train MSE: 1703605.95, Validation MSE: 2767148.70 Train RMSE: 1305.22, Validation RMSE: 1663.47 Train MAE: 1114.87, Validation MAE: 1455.68 Baseline R²: -0.01, Baseline MSE: 13799768.58, Baseline RMSE: 3714.80, Baseline MAE: 3263.89 Theoretical Prediction for 2024: [69390.93676949]
Row 4 - Port Value: Anchorage (AK) Row 4's Model - Train R²: 0.68, Validation R²: 0.70 Train MSE: 117597.29, Validation MSE: 104211.18 Train RMSE: 342.92, Validation RMSE: 322.82 Train MAE: 262.41, Validation MAE: 245.46 Baseline R²: -0.00, Baseline MSE: 349361.53, Baseline RMSE: 591.07, Baseline MAE: 483.67 Theoretical Prediction for 2024: [22730.78290248]
Row 5 - Port Value: Ashtabula (OH) Row 5's Model - Train R²: 0.48, Validation R²: 0.66 Train MSE: 9181267.33, Validation MSE: 5470474.22 Train RMSE: 3030.06, Validation RMSE: 2338.90 Train MAE: 2354.25, Validation MAE: 1916.85 Baseline R²: -0.00, Baseline MSE: 15988455.33, Baseline RMSE: 3998.56, Baseline MAE: 3484.76 Theoretical Prediction for 2024: [132953.11638069]
Row 6 - Port Value: Baltimore (MD) Row 6's Model - Train R²: 0.46, Validation R²: 0.53 Train MSE: 25876140.65, Validation MSE: 20372504.46 Train RMSE: 5086.86, Validation RMSE: 4513.59 Train MAE: 4211.21, Validation MAE: 3564.66 Baseline R²: -0.00, Baseline MSE: 43200535.39, Baseline RMSE: 6572.71, Baseline MAE: 5174.94 Theoretical Prediction for 2024: [81350.95571601]
Row 7 - Port Value: Barbers Point (HI) Row 7's Model - Train R²: 0.70, Validation R²: 0.68 Train MSE: 714840.36, Validation MSE: 576447.28 Train RMSE: 845.48, Validation RMSE: 759.24 Train MAE: 706.33, Validation MAE: 651.49 Baseline R²: -0.09, Baseline MSE: 1944110.40, Baseline RMSE: 1394.31, Baseline MAE: 1082.44 Theoretical Prediction for 2024: [-16397.88445091]
Row 8 - Port Value: Baton Rouge (LA) Row 8's Model - Train R²: 0.66, Validation R²: 0.68 Train MSE: 32311480.34, Validation MSE: 30557984.69 Train RMSE: 5684.32, Validation RMSE: 5527.93 Train MAE: 4889.70, Validation MAE: 4926.74 Baseline R²: -0.00, Baseline MSE: 94979746.30, Baseline RMSE: 9745.76, Baseline MAE: 8244.97 Theoretical Prediction for 2024: [-140388.47880554]
Row 9 - Port Value: Beaumont (TX) Row 9's Model - Train R²: 0.80, Validation R²: 0.80 Train MSE: 26025518.04, Validation MSE: 31587684.26 Train RMSE: 5101.52, Validation RMSE: 5620.29 Train MAE: 3785.21, Validation MAE: 3999.64 Baseline R²: -0.01, Baseline MSE: 162887181.83, Baseline RMSE: 12762.73, Baseline MAE: 9407.24 Theoretical Prediction for 2024: [1276929.58303833]
Row 10 - Port Value: Bellingham (WA) Row 10's Model - Train R²: 0.52, Validation R²: 0.73 Train MSE: 68361.09, Validation MSE: 51655.57 Train RMSE: 261.46, Validation RMSE: 227.28 Train MAE: 222.26, Validation MAE: 184.68 Baseline R²: -0.01, Baseline MSE: 193755.39, Baseline RMSE: 440.18, Baseline MAE: 342.55 Theoretical Prediction for 2024: [-13426.74582028]
Row 11 - Port Value: Boston(MA) Row 11's Model - Train R²: 0.65, Validation R²: 0.70 Train MSE: 3593923.43, Validation MSE: 2811203.65 Train RMSE: 1895.76, Validation RMSE: 1676.66 Train MAE: 1530.42, Validation MAE: 1281.26 Baseline R²: -0.01, Baseline MSE: 9557550.77, Baseline RMSE: 3091.53, Baseline MAE: 2610.78 Theoretical Prediction for 2024: [64950.48489761]
Row 12 - Port Value: Bridgeport (CT) Row 12's Model - Train R²: 0.65, Validation R²: 0.64 Train MSE: 58768.62, Validation MSE: 38657.26 Train RMSE: 242.42, Validation RMSE: 196.61 Train MAE: 180.01, Validation MAE: 150.86 Baseline R²: -0.00, Baseline MSE: 108570.44, Baseline RMSE: 329.50, Baseline MAE: 258.29 Theoretical Prediction for 2024: [26177.87896395]
Row 13 - Port Value: Brownsville (TX) Row 13's Model - Train R²: 0.67, Validation R²: 0.59 Train MSE: 157176.21, Validation MSE: 155344.76 Train RMSE: 396.45, Validation RMSE: 394.14 Train MAE: 309.56, Validation MAE: 291.25 Baseline R²: -0.01, Baseline MSE: 376900.14, Baseline RMSE: 613.92, Baseline MAE: 494.56 Theoretical Prediction for 2024: [23517.88686514]
Row 14 - Port Value: Charleston (SC) Row 14's Model - Train R²: 0.40, Validation R²: 0.06 Train MSE: 4844237.91, Validation MSE: 8078660.03 Train RMSE: 2200.96, Validation RMSE: 2842.30 Train MAE: 1595.48, Validation MAE: 1988.25 Baseline R²: -0.27, Baseline MSE: 10956720.60, Baseline RMSE: 3310.09, Baseline MAE: 2860.14 Theoretical Prediction for 2024: [-140780.605793]
Row 15 - Port Value: Christiansted (VI) Row 15's Model - Train R²: 0.59, Validation R²: 0.54 Train MSE: 49984336.58, Validation MSE: 43863708.64 Train RMSE: 7069.96, Validation RMSE: 6622.97 Train MAE: 4906.87, Validation MAE: 4386.43 Baseline R²: -0.08, Baseline MSE: 102763106.78, Baseline RMSE: 10137.21, Baseline MAE: 9351.62 Theoretical Prediction for 2024: [608856.2620163]
Row 16 - Port Value: Detroit (MI) Row 16's Model - Train R²: 0.73, Validation R²: 0.34 Train MSE: 196463538.63, Validation MSE: 340419530.52 Train RMSE: 14016.55, Validation RMSE: 18450.46 Train MAE: 9733.73, Validation MAE: 11129.33 Baseline R²: -0.05, Baseline MSE: 539832409.96, Baseline RMSE: 23234.29, Baseline MAE: 21103.60 Theoretical Prediction for 2024: [722017.35684967]
Row 17 - Port Value: Drummond Island (MI) Row 17's Model - Train R²: 0.47, Validation R²: 0.57 Train MSE: 28788754.80, Validation MSE: 25651850.42 Train RMSE: 5365.52, Validation RMSE: 5064.77 Train MAE: 3718.88, Validation MAE: 3446.03 Baseline R²: -0.01, Baseline MSE: 60163409.01, Baseline RMSE: 7756.51, Baseline MAE: 6643.76 Theoretical Prediction for 2024: [312415.85389709]
Row 18 - Port Value: Duluth/Superior (MI/WI) Row 18's Model - Train R²: 0.54, Validation R²: 0.50 Train MSE: 132989599.91, Validation MSE: 154226434.84 Train RMSE: 11532.11, Validation RMSE: 12418.79 Train MAE: 8720.82, Validation MAE: 9258.03 Baseline R²: -0.01, Baseline MSE: 310456008.07, Baseline RMSE: 17619.76, Baseline MAE: 16898.98 Theoretical Prediction for 2024: [-845215.24963379]
Row 19 - Port Value: Freeport (TX) Row 19's Model - Train R²: 0.49, Validation R²: 0.78 Train MSE: 28925110.03, Validation MSE: 9817129.46 Train RMSE: 5378.21, Validation RMSE: 3133.23 Train MAE: 3771.78, Validation MAE: 2445.58 Baseline R²: -0.08, Baseline MSE: 47999739.72, Baseline RMSE: 6928.18, Baseline MAE: 6650.51 Theoretical Prediction for 2024: [-271093.13784409]
Row 20 - Port Value: Galveston (TX) Row 20's Model - Train R²: 0.59, Validation R²: 0.28 Train MSE: 11700452.72, Validation MSE: 20357089.45 Train RMSE: 3420.59, Validation RMSE: 4511.88 Train MAE: 2281.13, Validation MAE: 2733.71 Baseline R²: -0.14, Baseline MSE: 31993898.35, Baseline RMSE: 5656.31, Baseline MAE: 4387.11 Theoretical Prediction for 2024: [-20643.40958261]
Row 21 - Port Value: Hampton Roads (VA) Row 21's Model - Train R²: 0.61, Validation R²: 0.69 Train MSE: 659413754.47, Validation MSE: 497181981.51 Train RMSE: 25679.05, Validation RMSE: 22297.58 Train MAE: 17989.06, Validation MAE: 14443.61 Baseline R²: -0.01, Baseline MSE: 1608369897.14, Baseline RMSE: 40104.49, Baseline MAE: 38598.70 Theoretical Prediction for 2024: [194977.04067993]
Row 22 - Port Value: Hilo (HI) Row 22's Model - Train R²: 0.59, Validation R²: 0.61 Train MSE: 65595.63, Validation MSE: 81345.10 Train RMSE: 256.12, Validation RMSE: 285.21 Train MAE: 186.85, Validation MAE: 219.88 Baseline R²: -0.00, Baseline MSE: 207920.26, Baseline RMSE: 455.98, Baseline MAE: 289.60 Theoretical Prediction for 2024: [1356.7988691]
Row 23 - Port Value: Hueneme (CA) Row 23's Model - Train R²: 0.65, Validation R²: 0.56 Train MSE: 9519850.46, Validation MSE: 16671452.11 Train RMSE: 3085.43, Validation RMSE: 4083.07 Train MAE: 1644.44, Validation MAE: 2266.73 Baseline R²: -0.04, Baseline MSE: 40024283.26, Baseline RMSE: 6326.47, Baseline MAE: 5056.08 Theoretical Prediction for 2024: [171627.0281105]
Row 24 - Port Value: Humboldt Bay (CA) Row 24's Model - Train R²: 0.21, Validation R²: 0.14 Train MSE: 231745.75, Validation MSE: 338809.53 Train RMSE: 481.40, Validation RMSE: 582.07 Train MAE: 380.05, Validation MAE: 469.26 Baseline R²: -0.01, Baseline MSE: 397546.00, Baseline RMSE: 630.51, Baseline MAE: 481.97 Theoretical Prediction for 2024: [14389.53980541]
Row 25 - Port Value: Huron (OH) Row 25's Model - Train R²: 0.35, Validation R²: 0.42 Train MSE: 24490270.70, Validation MSE: 24424560.39 Train RMSE: 4948.76, Validation RMSE: 4942.12 Train MAE: 4200.24, Validation MAE: 4188.24 Baseline R²: -0.03, Baseline MSE: 43732958.60, Baseline RMSE: 6613.09, Baseline MAE: 5468.72 Theoretical Prediction for 2024: [-62266.970191]
Row 26 - Port Value: Jacksonville (FL) Row 26's Model - Train R²: 0.33, Validation R²: 0.17 Train MSE: 44954702.82, Validation MSE: 66032696.61 Train RMSE: 6704.83, Validation RMSE: 8126.05 Train MAE: 5872.56, Validation MAE: 7053.86 Baseline R²: -0.04, Baseline MSE: 82713718.09, Baseline RMSE: 9094.71, Baseline MAE: 8224.63 Theoretical Prediction for 2024: [-192678.25743103]
Row 27 - Port Value: Lorain (OH) Row 27's Model - Train R²: 0.69, Validation R²: 0.75 Train MSE: 115701999.95, Validation MSE: 108625104.00 Train RMSE: 10756.49, Validation RMSE: 10422.34 Train MAE: 8730.33, Validation MAE: 9565.06 Baseline R²: -0.01, Baseline MSE: 438651599.33, Baseline RMSE: 20944.01, Baseline MAE: 19677.02 Theoretical Prediction for 2024: [953294.39849854]
Row 28 - Port Value: Marine City (MI) Row 28's Model - Train R²: 0.87, Validation R²: 0.73 Train MSE: 113166.67, Validation MSE: 307668.27 Train RMSE: 336.40, Validation RMSE: 554.68 Train MAE: 229.36, Validation MAE: 378.69 Baseline R²: -0.00, Baseline MSE: 1154563.39, Baseline RMSE: 1074.51, Baseline MAE: 1006.54 Theoretical Prediction for 2024: [28098.64863634]
Row 29 - Port Value: Monroe (MI) Row 29's Model - Train R²: 0.21, Validation R²: 0.37 Train MSE: 587232.32, Validation MSE: 756591.89 Train RMSE: 766.31, Validation RMSE: 869.82 Train MAE: 509.72, Validation MAE: 538.10 Baseline R²: -0.01, Baseline MSE: 1220534.67, Baseline RMSE: 1104.78, Baseline MAE: 739.28 Theoretical Prediction for 2024: [-49426.27476978]
Row 30 - Port Value: Muskegon (MI) Row 30's Model - Train R²: 0.38, Validation R²: 0.22 Train MSE: 7041025.88, Validation MSE: 9181061.29 Train RMSE: 2653.49, Validation RMSE: 3030.03 Train MAE: 2245.82, Validation MAE: 2708.33 Baseline R²: -0.05, Baseline MSE: 12294049.70, Baseline RMSE: 3506.29, Baseline MAE: 3138.68 Theoretical Prediction for 2024: [64121.05968857]
Row 31 - Port Value: Nawiliwili (HI) Row 31's Model - Train R²: 0.45, Validation R²: 0.66 Train MSE: 57187.88, Validation MSE: 47672.54 Train RMSE: 239.14, Validation RMSE: 218.34 Train MAE: 182.72, Validation MAE: 179.02 Baseline R²: -0.00, Baseline MSE: 141604.41, Baseline RMSE: 376.30, Baseline MAE: 294.97 Theoretical Prediction for 2024: [21491.81038547]
Row 32 - Port Value: New London (CT) Row 32's Model - Train R²: 0.18, Validation R²: 0.08 Train MSE: 6159322.21, Validation MSE: 7124465.66 Train RMSE: 2481.80, Validation RMSE: 2669.17 Train MAE: 2047.05, Validation MAE: 2252.57 Baseline R²: -0.03, Baseline MSE: 7981954.44, Baseline RMSE: 2825.24, Baseline MAE: 2511.59 Theoretical Prediction for 2024: [2189.80269408]
Row 33 - Port Value: Olympia (WA) Row 33's Model - Train R²: 0.45, Validation R²: 0.34 Train MSE: 9794882.70, Validation MSE: 10596596.69 Train RMSE: 3129.68, Validation RMSE: 3255.24 Train MAE: 2603.94, Validation MAE: 2728.10 Baseline R²: -0.04, Baseline MSE: 16589372.93, Baseline RMSE: 4073.01, Baseline MAE: 3603.95 Theoretical Prediction for 2024: [-168589.74404907]
Row 34 - Port Value: Orange (TX) Row 34's Model - Train R²: 0.77, Validation R²: 0.82 Train MSE: 31501288.85, Validation MSE: 28111629.02 Train RMSE: 5612.60, Validation RMSE: 5302.04 Train MAE: 3715.15, Validation MAE: 3709.07 Baseline R²: -0.03, Baseline MSE: 157470707.60, Baseline RMSE: 12548.73, Baseline MAE: 10936.63 Theoretical Prediction for 2024: [59488.01167679]
Row 35 - Port Value: Oswego (NY) Row 35's Model - Train R²: 0.62, Validation R²: 0.65 Train MSE: 330121973.96, Validation MSE: 306813769.25 Train RMSE: 18169.26, Validation RMSE: 17516.10 Train MAE: 14539.18, Validation MAE: 13448.28 Baseline R²: -0.01, Baseline MSE: 887321172.46, Baseline RMSE: 29787.94, Baseline MAE: 23300.99 Theoretical Prediction for 2024: [2250010.54302979]
Row 36 - Port Value: Palm Beach (FL) Row 36's Model - Train R²: 0.80, Validation R²: 0.76 Train MSE: 452988.12, Validation MSE: 431940.70 Train RMSE: 673.04, Validation RMSE: 657.22 Train MAE: 489.41, Validation MAE: 517.77 Baseline R²: -0.00, Baseline MSE: 1815576.21, Baseline RMSE: 1347.43, Baseline MAE: 1181.71 Theoretical Prediction for 2024: [-82218.15686035]
Row 37 - Port Value: Pascagoula (MS) Row 37's Model - Train R²: 0.66, Validation R²: 0.69 Train MSE: 97466340.72, Validation MSE: 67484981.76 Train RMSE: 9872.50, Validation RMSE: 8214.92 Train MAE: 7479.83, Validation MAE: 6759.03 Baseline R²: -0.00, Baseline MSE: 219015462.01, Baseline RMSE: 14799.17, Baseline MAE: 13685.44 Theoretical Prediction for 2024: [266962.31147766]
Row 38 - Port Value: Philadelphia (PA) Row 38's Model - Train R²: 0.37, Validation R²: 0.17 Train MSE: 549685.95, Validation MSE: 676420.00 Train RMSE: 741.41, Validation RMSE: 822.45 Train MAE: 540.67, Validation MAE: 566.91 Baseline R²: -0.03, Baseline MSE: 840479.24, Baseline RMSE: 916.78, Baseline MAE: 736.86 Theoretical Prediction for 2024: [-9994.26999784]
Row 39 - Port Value: Plaquemines (LA) Row 39's Model - Train R²: 0.69, Validation R²: 0.64 Train MSE: 84740167.67, Validation MSE: 95122211.05 Train RMSE: 9205.44, Validation RMSE: 9753.06 Train MAE: 6794.27, Validation MAE: 7527.46 Baseline R²: -0.06, Baseline MSE: 278581749.11, Baseline RMSE: 16690.77, Baseline MAE: 13943.80 Theoretical Prediction for 2024: [-838993.44750977]
Row 40 - Port Value: Ponce (PR) Row 40's Model - Train R²: 0.29, Validation R²: 0.30 Train MSE: 12608227.69, Validation MSE: 10437096.59 Train RMSE: 3550.81, Validation RMSE: 3230.65 Train MAE: 2635.04, Validation MAE: 2384.69 Baseline R²: -0.07, Baseline MSE: 15874744.46, Baseline RMSE: 3984.31, Baseline MAE: 3787.70 Theoretical Prediction for 2024: [64787.60416794]